perClass Documentation
version 5.1 (31-May-2017)

kb1: How to make decisions at a default operating point?

Keywords: setting operation point

Problem: How to make decision at a default operating point without ROC analysis.

Solution: Use sddecide directly on the trained pipeline.

>> p        %  trained pipeline
Gaussian model pipeline 2x3  3 classes, 3 components (sdp_normal)
>> pd=sddecide(p)
sequential pipeline     2x1 'Gaussian model+Decision'
 1  Gaussian model          2x3  3 classes, 3 components (sdp_normal)
 2  Decision                3x1  weighting, 3 classes, 1 ops at op 1 (sdp_decide)
>> dec=b(1:5,:)*pd 
sdlab with 5 entries, 2 groups: 'apple'(3) 'banana'(2)
>>+dec

apple 
apple 
apple 
banana
banana 

The pipeline pd is now able to make decisions. Therefore, if applied to a new data set it returns the classifier decision.

Ulternatively, the * symbol maybe used.

>> dec=b(5:10,:)*(p*sddecide); +dec

banana
banana
stone 
stone 
banana
stone